SEO in 2026: what really works in the age of AI

If there’s one thing 2026 is teaching us, it’s that search isn’t what it used to be. Not because Google has disappeared – quite the opposite, in fact – but because the way people search has changed radically.
Today, a query might start on ChatGPT, continue on Google, and perhaps end up on a website, review, or video. And along this journey, what really changes is not the channel, but the expectation: users no longer want to explore; they want to understand immediately.
This has led many to wonder whether SEO still makes sense. The answer is yes, but with one important caveat: SEO in 2026 is no longer about rankings; it is about relevance.

Recently, there has been much talk of artificial intelligence as a potential ‘replacement’ for search engines. In reality, the opposite is happening. AI does not eliminate search; it enhances it. And above all, it feeds off the same ecosystem: the web.
This means that every answer generated by an AI system still originates from content published online. If your website lacks accessibility, structure, and credibility, it won't be considered. Neither by Google nor by an AI assistant.
Italian data also confirms this trend. AI overviews are appearing in an increasing proportion of searches, and, often, users are already finding the answer they’re looking for there. This trend has led to a drop in organic clicks, but not to a reduction in the importance of SEO. On the contrary: today, being visible means being part of the answer, not just part of the SERP.

People's search habits also reflect this change. For years, SEO has been based on keywords: finding the right words, placing them in the right spots, and climbing the rankings.
Today, that is no longer enough.
A user doesn’t simply search for ‘hotel in Milan city centre’. More often, they ask something like, 'What is the best hotel in Milan for working, with reliable Wi-Fi?’. It is a more complex, more specific search, closer to a real need.
And here a fundamental difference emerges: those who continue to optimise for keywords are only scratching the surface. Those who, on the other hand, build content around users’ real problems become part of the decision-making process.
This applies to all sectors. An e-commerce site no longer competes solely on ‘men's running shoes' but on content that helps users choose, understand and avoid mistakes. A law firm should not publish dozens of similar articles but build a comprehensive and authoritative guide that becomes a point of reference.

In this scenario, the way we structure websites is also changing. Stand-alone pages are becoming less and less effective. Search engines and AI operate within a contextual framework: they look for consistency, depth and connections between pieces of content.
This is where thematic clusters come into play. Thematic clusters serve not as an 'advanced' SEO technique, but rather as a fundamental prerequisite for relevance. Content works when it is part of a system: linked to other content, supported by a clear structure, and part of a coherent ecosystem.

Another obvious change concerns content production. For years, people believed that publishing more was the right strategy. Today, the opposite is often true.
Artificial intelligence has made it incredibly easy to produce mediocre content. And precisely for this reason, its value has plummeted. Google has made its position clear through increasingly strict policies against mass-generated content that serves no real purpose.
In practice, the web no longer needs ‘more content’. It needs better content.
This means fewer articles, but more carefully crafted ones. Fewer pages, but more useful ones. Less volume, more substance.

One of the most intriguing aspects of SEO in 2026 is the emergence of a new implicit metric: citability.
When an AI system constructs a response, it must choose which sources to use. It does not simply select the most optimised pages; it selects those that are easy to understand, extract, and summarise.
AI systems are more likely to cite content that features clear definitions, logical structures, verifiable data, and signs of reliability. In this sense, writing elegantly is no longer just a matter of style but of visibility.

All of this leads to another realisation: SEO is no longer a silo.
For years, it was treated as a separate channel, with its KPIs and independent logic. Today, this approach is no longer effective. Visibility stems from interactions between multiple touchpoints: search, AI, social media, content and branding.
A user might discover a company via ChatGPT, search for it on Google, read a review and then return days later to convert. SEO is present in this journey, but it is never the only factor.
Such behaviour makes measurements more complex but no less relevant. Indeed, it forces us to think more realistically: the value of SEO lies not in traffic but in its impact on the business.

Finally, a recurring question: will websites still be needed?

The answer is yes, but in a different role. A website is no longer just a destination for users; it is a source for search ecosystems. It is the place where the official, structured version of the information that AI systems will use resides.

For this reason, factors such as structure, performance and accessibility are not technical details but rather prerequisites. Without these foundations, you won’t even be in the running.

SEO in 2026 is therefore very different from what we used to know. Not because the rules have changed completely, but because the context in which they apply has changed.
The focus has shifted from optimising pages for search engines to creating content that deserves selection by both users and AI.
And ultimately, the principle remains surprisingly simple: the winner is whoever provides the best answer.